AMD phenom makes a comeback?

[Cronus]

WHY DO YOU DESTROY MY DREAMS MAN?! D':

 

You're a professional bird, your dreams can't go anywhere but south! :lol:

[Bluejay0]

You're a professional bird, your dreams can't go anywhere but south! :lol:

You sir. Have won my respect.

[Inkz]

Would be great if this was true, but I feel like it is too good to be true.

[Dravic]

Same as with 'cpuboss' claiming 9970 having 8TFlops etc. :D

 

In other words:

 

Impossible bullshit. Nothing to see here, move along.

[AdamG1401]

I can see 12 cores but 6 GHz is daft. AMD should just filling cores at it rather than giving us these CPU's that have insane TDP/voltage requirements at unreal prices.

[techswede]

i want to believe! :( if they start selling phenoms again i want to build a second rig with it!
a 6ghz 12 core might just be able to kick intel in the teeth hard enough to make them get their act together (ivy + haswell temp issues)

[Dravic]

http://linustechtips.com/main/topic/42791-phenom-ii-x6-smokes-many-intel-and-all-amd-cpus-in-dirt-showdown-d/

 

Dirt Showdown is an anomaly, I still wonder how's that possible :D

[BernardoSLR]

That CPU-Z screen... Such a blatant fake! XD

[Esivo]

And I was thinking of building a new PC. If this is true, I need to get back to the drawing board. 

 

Awaiting confirmation.

 

Please be true *fingers crossed*

[DLM_012]

I want to believe.

 

I don't see how this is possible though based off of my (limited) knowledge of computer hardware. 6Ghz and 75W TDP sounds impossible to do with one chip. What if AMD has been yanking our chains all along and been working on this beast? That would be some shocker. I have no plans to upgrade my 8350 but if this is legit, I gotta be a part of history and buy this thing.

[Opsilon]

This would be awesome, but i really doubt it.

6ghz on 12 cores and 75w TDP with 1.4 V is kinda strange

I mean lets be realistic, 12 cores today on a consumer product is not really needed ( intel extreme  CPU-s are 6 cores even with 12 threads )

25nm processes are fine but 75 w TDP? that is just not possible. Sure CPU-z does not have this cpu in the database, so it might be a wrong readout.

 

So on a realistic side this should be the 12 cores around 5 ghz PC with a 150+ TDP, and if this is true this would require a different socket AM3+ would not have been enough, but AMD claims no new sockets for 2014, and 1.4 v is a high voltage, so i would think this is not really the case, i mean intel haswell had leaks with like 0.8V and over 4.6 ghz and i have not seen a chip like that ever. I am aware of 12 cores and a high frequency needs more volts, but today 1.4 is too much, and since the die shrink, this voltage is an overkill, the point of shrinking the process is to lower volts and heat, and ad more transistors.

 

I would love for this to be the new steamroller, but after the original ivy bridge launch i lost my hope, since we are closing to the end of silicon based chips, we will not see a huge improvement every generation, maybe more features, but not a good IPC boost, i think sandy bridge did that last, and we will not see anything like that very soon.

 

The other thing is that this might be a new AMD server, and have no connection to Phenoms or FX chips...

[SubTract]

I noticed a few things really quick...

1. Like @qwertywarrior pointed out they write at the end that its based on false assumptions by the person who leaked it.

2. Phenom IV.... Where did Phenom III go?

3. 6GHz

4. 75w TPD

5. Didn't AMD's roadmap show them going from 28nm to 20nm

 

A lot of people here seem to think that 6Ghz on 12 cores with a 75w TDP is impossible...

All I have to say to you is that it was once thought that nobody would fill up.

Bill Gates once said "Nobody will ever need more than 640k RAM"

They also moved to multi core CPU's because they were having problem with getting more than 4Ghz.

 

it wouldn't surprise me if it happened in the next... decade...

[Dravic]

I noticed a few things really quick...

1. Like @qwertywarrior pointed out they write at the end that its based on false assumptions by the person who leaked it.

2. Phenom IV.... Where did Phenom III go?

3. 6GHz

4. 75w TPD

5. Didn't AMD's roadmap show them going from 28nm to 20nm

 

A lot of people here seem to think that 6Ghz on 12 cores with a 75w TDP is impossible...

All I have to say to you is that it was once thought that nobody would fill up.

Bill Gates once said "Nobody will ever need more than 640k RAM"

They also moved to multi core CPU's because they were having problem with getting more than 4Ghz.

 

it wouldn't surprise me if it happened in the next... decade...

 

Nope, it will never happen because silicion is at its limits already - or almost.

 

We can only squish extra performance by making better architecture and jumping to smaller nodes, nothing more.

 

And we're already at 14nm. I guess 10nm will be as good as we can get until 2030 and until 2030 i hope for graphene to be commonly used which will be superior to silicion.

 

In other words, we wont gain more than +50% performance over next 10 years with silicon compared to i7 4770 (single core performance) but we may gain it by switching to graphene once it's sorted out - we currently do not have technology to detect 0 and 1 states on graphene so it's pointless, can't make any computing with it.

[qwertywarrior]

Nope, it will never happen because silicion is at its limits already - or almost.

We can only squish extra performance by making better architecture and jumping to smaller nodes, nothing more.

And we're already at 14nm. I guess 10nm will be as good as we can get until 2030 and until 2030 i hope for graphene to be commonly used which will be superior to silicion.

In other words, we wont gain more than +50% performance over next 10 years with silicon compared to i7 4770 (single core performance) but we may gain it by switching to graphene once it's sorted out - we currently do not have technology to detect 0 and 1 states on graphene so it's pointless, can't make any computing with it.

Continuing the pattern of halving the surface area of each node (or reducing the square sides by 1.4) every two years, it can be expected that 4 nm technology will appear in roughly 2022, 3 nm in 2024, 2 nm in 2026, and lastly 1 nm in 2028, which may be the limit for current lithographic semiconductor silicon technology as presently implemented .As a result 2030 and beyond may see a shift in technology to enable smaller components in the form of silicene or silicon nanotubes or a move towards other processing methods such as quantum computing.

-wikipedia

there was also a news article like 1 week ago about some kind of conference where they were talking about the materials that would be used with silicon 7 nm and 5 nm

[SubTract]

Nope, it will never happen because silicion is at its limits already - or almost.

 

We can only squish extra performance by making better architecture and jumping to smaller nodes, nothing more.

 

And we're already at 14nm. I guess 10nm will be as good as we can get until 2030 and until 2030 i hope for graphene to be commonly used which will be superior to silicion.

 

In other words, we wont gain more than +50% performance over next 10 years with silicon compared to i7 4770 (single core performance) but we may gain it by switching to graphene once it's sorted out - we currently do not have technology to detect 0 and 1 states on graphene so it's pointless, can't make any computing with it.

 

They can already get those frequencies on silicone... Albeit with extreme cooling. Its all about efficiency. There is always room for improvement.

Who knows what the actual limit of silicone is. But its pretty easy to say that we don't actually know. A lot of things have been "Impossible" only to be made possible down the line.

I wouldn't be so quick to jump to a conclusions. 

 

Who knows we might move to something else. Maybe not graphene just yet but if they actually find they're limited by silicone and they can't do anything more with it then they'll put the effort into another base. 

 

The solution could also be physically larger chips. Can't fit more stuff into a tiny area, Add more area. 

[Dravic]

Continuing the pattern of halving the surface area of each node (or reducing the square sides by 1.4) every two years, it can be expected that 4 nm technology will appear in roughly 2022, 3 nm in 2024, 2 nm in 2026, and lastly 1 nm in 2028, which may be the limit for current lithographic semiconductor silicon technology as presently implemented .As a result 2030 and beyond may see a shift in technology to enable smaller components in the form of silicene or silicon nanotubes or a move towards other processing methods such as quantum computing.

-wikipedia

there was also a news article like 1 week ago about some kind of conference where they were talking about the materials that would be used with silicon 7 nm and 5 nm

 

I recommend Michio Kaku's videos and points here, for example:

 

 

Lots of videos from him about computing and physics etc.

[Opsilon]

All good facts guys and i love how this has not become an all out fanboy war XD

 

You are right we will not see this, and yeah where the hell is phenom III gone of to?

But anyway it's a fun topic, for digging information, and as far as i know they can potentially reduce the the transistors to molecular size, maybe an atom, but we ( so far ) can't get even there.

It will be more cost beneficial to find a new way, and switch to a new material, but that is a story for another time anyway.

 

I potentially see this as an FX chip like i stated earlier, but not steamroller, but excavator instead, as a high end model, that will require a new socket, and that we will not see mainstream 12 core AMD chip at all on AM3+, and even if i am right i do not see this until late 2015 or even later. The new consoles will launch this year, i doubt AMD will push for more cores just yet except if this is a new socket AMD enthusiast class CPU.

Not to mention AMD should really try and improve the IPC of their cores and than increase the count, i mean if a phenom II core can beat an fx8100 core, that is very sad for a newer chip.

 

On the other hand if this is coming soon we will finally see a competitive AMD on the high end

[d3nia1]

It's fake... just look at the comments on the "article."

[MeltyMoon]

Guys....25nm is for NAND flash, this is a hoax  :ph34r:

[_ASSASSIN_]

Guys....25nm is for NAND flash, this is a hoax  :ph34r:

good to know the AMD fanboy isn't jumping up and down in excitment

[Dravic]

All good facts guys and i love how this has not become an all out fanboy war XD

 

You are right we will not see this, and yeah where the hell is phenom III gone of to?

But anyway it's a fun topic, for digging information, and as far as i know they can potentially reduce the the transistors to molecular size, maybe an atom, but we ( so far ) can't get even there.

It will be more cost beneficial to find a new way, and switch to a new material, but that is a story for another time anyway.

 

I potentially see this as an FX chip like i stated earlier, but not steamroller, but excavator instead, as a high end model, that will require a new socket, and that we will not see mainstream 12 core AMD chip at all on AM3+, and even if i am right i do not see this until late 2015 or even later. The new consoles will launch this year, i doubt AMD will push for more cores just yet except if this is a new socket AMD enthusiast class CPU.

Not to mention AMD should really try and improve the IPC of their cores and than increase the count, i mean if a phenom II core can beat an fx8100 core, that is very sad for a newer chip.

 

On the other hand if this is coming soon we will finally see a competitive AMD on the high end

 

You are very optimistic... Phenom II x6 beat FX-8350 in most tasks when clock-to-clock compared. Can you imagine? :D

 

Phenom II x6 just got much better IPC. They didnt go higher than 4,0Ghz, really, and that was the problem with them. If they could reach 5,0GHz like FX series, we would have a chip that can fight with Sandy Bridge in single thread.

[YellowDragon]

I call bullocks, You wouldn't waste billions developing a new architecture to produce newer CPU and then just decide to backstep, If AMD would have taken a step back it would have been after the bulldozer launch, considering that they are on the verge of another CPU refresh based off the same design mentality as bulldozer and vishera it does not make a single bit of sense. 

[greenroost1445]

Guys....25nm is for NAND flash, this is a hoax  :ph34r:

Well so much for that...  :angry:

[Vic9mm]

You are very optimistic... Phenom II x6 beat FX-8350 in most tasks when clock-to-clock compared. Can you imagine? :D

 

Phenom II x6 just got much better IPC. They didnt go higher than 4,0Ghz, really, and that was the problem with them. If they could reach 5,0GHz like FX series, we would have a chip that can fight with Sandy Bridge in single thread.

Maybe so but when it comes to gaming my 1100T was bottlenecking the hell out of my GTX 770LE. When I bought the FX-8350 my FPS when up about 35-40%

[Dravic]

Maybe so but when it comes to gaming my 1100T was bottlenecking the hell out of my GTX 770LE. When I bought the FX-8350 my FPS when up about 35-40%

 

Did you overclock 1100T to 4GHz and boosted the north bridge etc as should be done or did you simply overclock with multiplier like other noobs? :D